|Publication number||US5252501 A|
|Application number||US 07/814,547|
|Publication date||Oct 12, 1993|
|Filing date||Dec 30, 1991|
|Priority date||Dec 30, 1991|
|Also published as||EP0550021A2, EP0550021A3|
|Publication number||07814547, 814547, US 5252501 A, US 5252501A, US-A-5252501, US5252501 A, US5252501A|
|Inventors||Mehrdad M. Moslehi|
|Original Assignee||Texas Instruments Incorporated|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (5), Referenced by (13), Classifications (20), Legal Events (7)|
|External Links: USPTO, USPTO Assignment, Espacenet|
Advanced CMOS and BiCMOS technologies employ twin (n and p) wells in order to allow simultaneous optimization of the NMOS and PMOS transistors. The conventional twin well formation processes employ one or two microlithography masking steps. The two-mask fabrication process has the advantage that it does not degrade the silicon surface planarity or topography. This is an important requirement in advanced sub-0.5 μm CMOS and BiCMOS technologies where even a small (e.g. a few thousand Å) step between the n-well and p-well regions can result in gate length variations between the NMOS and PMOS transistors (due to the limited depth-of-focus orDOF budget in advanced photo-lithography tools). The CMOS transistor gate length variations (after patterning and etch) can degrade manufacturability and yield. As a result, the process simplicity of the conventional one-mask fabrication process is not a strong and sufficient advantage/improvement to justify its use in sub-0.5 μm semiconductor technologies since the one-mask process consumes silicon when masking a particular area to provide a well region. Such oxidation-induced silicon consumption results in uneven planarity of the PMOS and NMOS gates as shall be explained below. The smaller the device gate-lengths the more critical uneven planarity becomes. As discussed, the potential gate length variability caused by the nonplanar n and p well surfaces and DOF limitation imposes the above-mentioned obstacle. The surface topography problem in the conventional one-mask fabrication process is generated by a selective thermal oxidation process. An oxide/nitride stack is patterned and is used (usually along with the photoresist mask) as an ion implantation mask to define one of the wells. After removal of the photoresist, an oxidation step is performed which selectively defines an oxide hard mask over the implanted region. A second ion implantation step is used to define the second (opposite) well regions. As this process flow indicates, a surface topography or step is produced between the n and p well regions due to the silicon consumption by the selective thermal oxidation step.
Thus, advanced sub-0.5 μm CMOS and BiCMOS technologies usually rely on a two-mask (non-self aligned) process to define the n-well and p-well regions. In the two-mask process, patterned photoresist (or hard mask) layers are used as ion implantation masks. No surface step or topography is produced, therefore allowing uniform PMOS and NMOS gate patterning. However, the use of one additional masking process adds to the overall process cost and complexity.
Therefore, there is a need for a simple self-aligned manufacturable single-mask process for formation of n-well and p-well regions without the surface topography and gate patterning DOF problems of the conventional one-mask methods.
FIGS. 1 through 5 illustrate cross-sectional drawings of the resulting structure from the following processing steps.
Applicable reference numerals have been carried forward.
A single-mask self-aligned process is disclosed for formation of n and p wells for advanced CMOS and BiCMOS technologies. The proposed process forms n-well and p-well regions using a single microlithography masking step along with a selective semiconductor (Si or GeSi) growth (SSG) process without producing surface topography or degrading surface planarity. This simple process ensures uniform NMOS and PMOS gate patterning due to flat surface topography. The n-to-p well placement is self-aligned due to the use of a disposable SSG hard mask.
The invention provides a single-mask self-aligned twin well formation process for advanced CMOS and BiCMOS technologies. The process is based on the use of a selectively grown semiconductor (or metal) hard mask. The process flow for the preferred embodiment of the process flow is as follows:
1. With reference to FIG. 1 which illustrates a portion of the resulting structure formed from process steps of the invention, fabricate a stack of oxide 2/polysilicon 4/oxide 6 over, for instance, a p epitaxial layer lying on a p+ (highly doped p-type) substrate using thermal oxidation, low pressure chemical-vapor deposition (LPCVD) of polysilicon, and LPCVD or plasma enhanced chemical-vapor deposition (PECVD) of oxide. Note that the p-epitaxial layer and the p+ substrate may be collectively called the substrate. Lower oxide layer 2 may be 100 to 1000 Å (e.g. 200 Å). Polysilicon layer 4, which acts as a buffer layer may be 200-1000 Å (e.g. 500 Å). Top oxide layer 6 is chosen to be 5000-15000 Å (e.g. 1 μm).
2. With reference to FIG. 2 which illustrates a cross-sectional drawing of the resulting structure from the following processing steps, use a well microlithography mask and appropriate photoresist to pattern and etch oxide 6 using an anisotropic oxide etch to form an oxide hard mask 7. Polysilicon layer 4 acts as an etch-stop layer for the oxide etch process. The exposed polysilicon regions from which oxide was removed correspond to one polarity of the well regions, for instance n-well regions, to be formed later in the p-epitaxial layer. Next perform a n-type (such as phosphorus) ion implantation, indicated by arrows 10, to incorporate the first well (e.g. n-well) dopant atoms into the substrate through the exposed thin polysilicon windows. The thick patterned oxide 7 will prevent dopant incorporation in its underlying substrate regions, thus acting as an ion implantation hard mask. Strip the photoresist. Note that the photoresist may be stripped before or after the well ion implantation. The resulting n-well implant profile is labeled 12 in FIG. 2.
3. With reference to FIG. 3 which illustrates a cross-sectional drawing of the resulting structure from the following processing steps, perform selective semiconductor growth (SSG) to form SSG layer 14 using the exposed polysilicon regions above implant 12 as seed layers for selective silicon (or silicon germanium) deposition. The selectivity of the growth should preferably be as high as possible so as to result in a well defined hard mask to be formed by the SSG layer as shall be explained below. Facets and sidewall or bulk defects will not cause any problem. This is due to the fact that SSG layer 14 will be used as a disposable hard mask for a well region to be formed of an opposite conductivity type from that discussed in step 2 during a subsequent well ion implanation. SSG layer 14 may be grown to be as thick as the oxide hard mask (e.g. 1 μm). Next, remove the thick oxide hard mask 7 shown in FIG. 2 by using a selective oxide etch such as a wet etch or a vapor-phase etch. Alternatively, a layer formed from selective deposition processes such as chemical-vapor deposited tungsten (CVD-W) may be used in place of the SSG layer. However, the SSG layer formation is a process step for the preferred embodiment of the invention.
4. With reference to FIG. 4 which illustrates a cross-sectional view of the resulting structure from the following processing steps form dielectric spacers 16 such as oxide spacers using conformal oxide deposition, e.g. low-pressure chemical-vapor deposition (LPCVD) using tetraethylorthosilicate (TEOS), and an anisotropic oxide etch-back process. Formation of spacers 16 is optional. Alternatively, nitride spacers can optionally be formed instead of oxide. The thickness of spacer 16 (e.g. 5000 Å) will determine the self-aligned spacing between the n-well and p-well implants (excluding the lateral implant straggle or spread). Optional spacers 16 serve to minimize the lateral dopant compensation due to intermixing of the n-well and p-well profiles after well diffusion or drive-in. Next, form a well of the opposite conductivity from the wells to be formed in the previous steps. For instance, assuming n-wells were to be formed previously, perform a p-type ion implantation (i.e. boron ion implantation), indicated by arrows labeled 18, overall. SSG regions 14 and dielectric spacers 16 will act as self-aligned hard masks during implantation. The resultant p-well implant profile is shown as region 15.
5. With reference to FIG. 5 which illustrates a cross-sectional drawing of the resulting structure from the following process steps, perform a selective etch, such as a wet etch or vapor-phase etch to remove spacers 16 (assuming oxide spacers) shown in FIG. 4. Next perform a selective polysilicon etch to clear the surface down to oxide layer 2. Diffuse the n-well and p-well regions using furnace or rapid thermal processing-based anneal in, for instance, an ammonia ambient (to take advantage of the oxynitridation-enhance diffusion or ONED effect) or in an oxidizing or inert ambient. N-wells and p-wells are thus formed.
6. The well regions now being appropriately formed, proceed with any remaining desired device fabrication steps to complete the device manufacturing process.
Although the invention has been described in detail herein with reference to its preferred embodiment and certain described alternatives, it is to be understood that this description is by way of example only, and it is not to be construed in a limiting sense. It is to be further understood that numerous changes in the details of the embodiments of the invention and additional embodiments of the invention, will be apparent to, and may be made by, persons of ordinary skill in the art having reference to this description. For instance, p-wells could have been formed in place of n-wells and vice versa. It is contemplated that all such changes and additional embodiments are within the spirit and true scope of the invention as claimed below.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4050965 *||Oct 21, 1975||Sep 27, 1977||The United States Of America As Represented By The Secretary Of The Air Force||Simultaneous fabrication of CMOS transistors and bipolar devices|
|US4424621 *||Dec 30, 1981||Jan 10, 1984||International Business Machines Corporation||Method to fabricate stud structure for self-aligned metallization|
|US4527325 *||Dec 23, 1983||Jul 9, 1985||International Business Machines Corporation||Process for fabricating semiconductor devices utilizing a protective film during high temperature annealing|
|US4558508 *||Oct 15, 1984||Dec 17, 1985||International Business Machines Corporation||Process of making dual well CMOS semiconductor structure with aligned field-dopings using single masking step|
|US5132241 *||Apr 15, 1991||Jul 21, 1992||Industrial Technology Research Institute||Method of manufacturing minimum counterdoping in twin well process|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5547894 *||Dec 21, 1995||Aug 20, 1996||International Business Machines Corporation||CMOS processing with low and high-current FETs|
|US5583062 *||Jun 7, 1995||Dec 10, 1996||Lsi Logic Corporation||Self-aligned twin well process having a SiO2 -polysilicon-SiO2 barrier mask|
|US5670393 *||Jul 12, 1995||Sep 23, 1997||Lsi Logic Corporation||Method of making combined metal oxide semiconductor and junction field effect transistor device|
|US5670395 *||Apr 29, 1996||Sep 23, 1997||Chartered Semiconductor Manufacturing Pte. Ltd.||Process for self-aligned twin wells without N-well and P-well height difference|
|US5763302 *||Aug 20, 1996||Jun 9, 1998||Lsi Logic Corporation||Self-aligned twin well process|
|US5770492 *||Dec 18, 1996||Jun 23, 1998||Lsi Logic Corporation||Self-aligned twin well process|
|US5913115 *||Apr 29, 1998||Jun 15, 1999||Siemens Aktiengesellschaft||Method for producing a CMOS circuit|
|US5956583 *||Jun 30, 1997||Sep 21, 1999||Fuller; Robert T.||Method for forming complementary wells and self-aligned trench with a single mask|
|US6207538||Dec 28, 1999||Mar 27, 2001||Taiwan Semiconductor Manufacturing Company||Method for forming n and p wells in a semiconductor substrate using a single masking step|
|US6235568 *||Jan 22, 1999||May 22, 2001||Intel Corporation||Semiconductor device having deposited silicon regions and a method of fabrication|
|US6307230 *||Oct 12, 1999||Oct 23, 2001||Texas Instruments Incorporated||Transistor having an improved sidewall gate structure and method of construction|
|US6586296 *||Apr 30, 2001||Jul 1, 2003||Cypress Semiconductor Corp.||Method of doping wells, channels, and gates of dual gate CMOS technology with reduced number of masks|
|US6593175 *||Mar 14, 2001||Jul 15, 2003||Advanced Micro Devices, Inc.||Method of controlling a shape of an oxide layer formed on a substrate|
|U.S. Classification||438/420, 438/526, 438/527, 257/E27.067, 257/E21.346, 257/E21.644, 257/371, 257/E27.015|
|International Classification||H01L21/266, H01L27/06, H01L21/265, H01L27/092, H01L21/8238, H01L21/8249|
|Cooperative Classification||H01L21/823892, H01L27/0623, H01L21/266, H01L27/0928|
|European Classification||H01L21/266, H01L21/8238W|
|Mar 2, 1992||AS||Assignment|
Owner name: TEXAS INSTRUMENTS INCORPORATED, TEXAS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:MOSLEHI, MEHRDAD M.;REEL/FRAME:006039/0743
Effective date: 19920214
|May 20, 1997||REMI||Maintenance fee reminder mailed|
|Jun 9, 1997||FPAY||Fee payment|
Year of fee payment: 4
|Jun 9, 1997||SULP||Surcharge for late payment|
|Mar 29, 2001||FPAY||Fee payment|
Year of fee payment: 8
|Mar 29, 2005||FPAY||Fee payment|
Year of fee payment: 12
|Apr 27, 2005||REMI||Maintenance fee reminder mailed|